On 2023/02/08 at 10:16 AM, lucass said:

in the aftenoon the voltage go down and it stack to 60v, purple line is the same solare panel of the 4 series connected to the inverter but it is connected to a victron mppt 100/20 

@mafiose13 do you have the same issue when the sun go up e when it go down?

I had one installation, that exhibited similarly an odd 60v base line during off-sun periods (see my previous post on this thread about it). Never went down to 0v at night. 

Since this was an update on a previous installation, changing inverter and installing newer panels 6x550 in series, the PV cables running from basement to roof of 6 story building (30meters) were re-used. It turns out that there was a missed junction box in which 20A blocking diodes were installed on those PV lines. Removing them removed that 60v night time baseline. I cannot explain why.

The diode flow orientation allowed current to flow from Panel positive cable to inverter and from inverter through negative cable to panels to complete that circuit in that flow,  but blocked any current in reverse. Not sure why they would be installed on a series panels and not sure why removing them removed the 60v. Maybe someone can explain it?

PS: switching off the panel breaker (while they were installed) gave 0v at night, that's how i narrowed it down to an issue with cabling/panel connections in the first place) 

@lucassNot sure if ur 60v is similar...but check ur lines, make sure nothing is blocking free flow in any direction or interfering  

Edited February 20, 20233 yr by mafiose13

1 hour ago, mafiose13 said:

I had one installation, that exhibited similarly an odd 60v base line during off-sun periods (see my previous post on this thread about it). Never went down to 0v at night. 

Since this was an update on a previous installation, changing inverter and installing newer panels 6x550 in series, the PV cables running from basement to roof of 6 story building (30meters) were re-used. It turns out that there was a missed junction box in which 20A blocking diodes were installed on those PV lines. Removing them removed that 60v night time baseline. I cannot explain why.

The diode flow orientation allowed current to flow from Panel positive cable to inverter and from inverter through negative cable to panels to complete that circuit in that flow,  but blocked any current in reverse. Not sure why they would be installed on a series panels and not sure why removing them removed the 60v. Maybe someone can explain it?

PS: switching off the panel breaker (while they were installed) gave 0v at night, that's how i narrowed it down to an issue with cabling/panel connections in the first place) 

@lucassNot sure if ur 60v is similar...but check ur lines, make sure nothing is blocking free flow in any direction or interfering  

Now i have 4 panel 550w VOC 50v but in cloudy condition the mppt doesn t work very well... sometimes it stop and the power go down to 0w

https://powerforum.co.za/topic/16007-axpert-vmiii-4kw-24v-firmware-miscellaneous/page/4/#comment-154334@mafiose13, please note!

This is patched firmware version 255.06, based on factory firmware 55.06 for the Axpert VM III non-twin, new models 4 kW or 6 kW. It is modelled on patched firmware version 656.05 for the Axpert Max II 10kW:

* Threshold current is reduced from 0.50 A to 0.10 A (656.05 used 0.05 A, but there was a comment saying that it may be too low).

* The premature float bug is fixed.

* Every minute, if the PV voltage is below 175 V, the target PV voltage is reset to 200 V and the direction is forced to 1, meaning that we expect the PV power to increase by increasing panel voltage.

* The QVFW command returns 255.06, meaning that the display still shows the U1 version to be 55 06 (representing 55.06), but monitoring software should report the version as 00255.06, possibly without the leading zeroes.

I believe that all VM models use the same DSP processor ('28062), so there should be no problem with chipsets. Do not use this patched firmware with any other model, including other Axpert VM IIIs with rated power other than 4.0 or 6.0 kW.

mafiose13 has reported that factory firmware version 55.06 fixed an issue he had with the AC output voltage not being set properly (ask for 230 V, get 215 V, for example). As far as I know, factory firmware version 55.06 is the latest firmware for non-twin Axpert VM IIIs (new type).

Firmware upload instructions.

Use at your own risk.

[ Edit: use of this patched firmware is deprecated for the time being; see this post. ]

[ Edit 2: Patch deleted; see patched firmware version 355.06 ]

 

 

 

Edited December 30, 20232 yr by Coulomb

@Coulomb

Many thanks for patching this version. 

I do have a couple of questions.

With reference to the Max II 56.05 thread, the 656.05 was the progressive best version after the many trials and error, and i take it the we put in the resulting lessons-learned into this one?

Was any of Georg patches ie shadow patch or any other included? 

Obviously aside from premature float, the other fixes specifically target stuck at 120v in cloudy/bad weather, no difference will be noticed in bad weather.

I've installed it on 1 unit, kept another for comparison on stock. I'll keep you posted on updates in couple days. 

 

4 hours ago, mafiose13 said:

the 656.05 was the progressive best version after the many trials and error, and i take it the we put in the resulting lessons-learned into this one?

Yes, that's the idea. Though I don't recall a lot of feedback after 656.05, so maybe there are better ways. More people testing will hopefully clarify the situation.

4 hours ago, mafiose13 said:

Was any of Georg patches ie shadow patch or any other included? 

No, only his current threshold, and I changed that from 0.05 A to 0.10 A.

4 hours ago, mafiose13 said:

the other fixes specifically target stuck at 120v in cloudy/bad weather, no difference will be noticed in bad weather.

? No difference will be detected with shading, but bad weather performance should be improved.

4 hours ago, mafiose13 said:

I'll keep you posted on updates in couple days. 

Thanks.

@Coulomb

First observation, there seems to be a glitch. It's a pretty sunny day today, however one passing small cloud briefly obstructed the sun for 15-20 seconds and went back to full sun. Obviously, power went terribly down but after that the MPPT stayed stuck at that mode voltage was 265-275....0.3-0.5A....50-150W. I waited 10 min of full sun and didn't recover. I had to manual switch off panel breaker and when on...it went back to 2.7kw production. 

Never done that before. 

 

Edit: The above was the first cloud of the day. 2hrs later, More clouds have come and go, this hasn't happended again for now. Odd! Maybe a learning curve?! 

I can already attest to the premature float bug being fixed.

Cloud have been intermittent, so can't judge that yet. 

Edited February 21, 20233 yr by mafiose13

5 hours ago, mafiose13 said:

Obviously, power went terribly down but after that the MPPT stayed stuck at that mode voltage was 265-275....0.3-0.5A....50-150W.

I think the problem is that this patch is calibrated for someone else's PV strings. The reset will only happen if the PV voltage drops below 175 V, and it will only reset to 200 V. That's a pretty narrow window.

Perhaps it did reset, and 265-275 V is above your maximum power point. When you get 2.7 kW, what voltage and current is that? Usually the problem is too low PV voltage; I suspect that in this case, the problem was too high PV voltage.

7 hours ago, Coulomb said:

I think the problem is that this patch is calibrated for someone else's PV strings. The reset will only happen if the PV voltage drops below 175 V, and it will only reset to 200 V. That's a pretty narrow window.

Perhaps it did reset, and 265-275 V is above your maximum power point. When you get 2.7 kW, what voltage and current is that? Usually the problem is too low PV voltage; I suspect that in this case, the problem was too high PV voltage.

At the peak production of 2.7-2.9kw, i am usually at 215-230v & 12.5-13.5A. 

I am running 6x550w Jinko tiger pro panels. (Specs below)

4 hours ago, mafiose13 said:

At the peak production of 2.7-2.9kw, i am usually at 215-230v & 12.5-13.5A. 

I am running 6x550w Jinko tiger pro panels. (Specs below)

Ok, thanks. So indeed, 270 V (45 V per panel) is above the Vmp voltage, and is heading towards Voc, hence the low current and power.

I hope it's a one-off, and we don't have to encourage the MPPT to look lower in voltage. Or patch in a full MPPT sweep algorithm.

Now i have the 56.11 firmaware, 4 panels in series so 160v and 200VOC, with this value my mppt doesn t start at the morning and doesn t stop at evening

 

 

19 hours ago, Coulomb said:

Ok, thanks. So indeed, 270 V (45 V per panel) is above the Vmp voltage, and is heading towards Voc, hence the low current and power.

I hope it's a one-off, and we don't have to encourage the MPPT to look lower in voltage. Or patch in a full MPPT sweep algorithm.

Today was purely sunny not a single cloud in sight, so nothing happened of the sort. I'll keep you posted on that issue.

However, i noticed another problem yesterday but didn't make much of it because i thought clouds were coming and going maybe it didn't reach float and never check battery charging voltage. But today with sunny weather I can confirm that we went from "premature float" to "never float". 🙂

We reached bulk target voltage 29.2v by 10AM, by 2PM charging amps were down to 0A charge at full bulk 29.2v and it stayed at 29.2v till sunset 5pm, it showed 0 to 4 amps charging during this time, while oddly discharging 0-4amps oppositely....but maintained bulk voltage 29.2v never switched to float voltage 27.2v. 

I have LA batteries. 

This has got to be Not normal. 

Edit: (11pm) Low battery 4 alarm, while battery at 24.8v. 🙃

I'm thinking the inverter is now measuring and thinking that 29.2v is float and referencing it in calculating SOC, thus an alarm at this voltage. 

 

 

Edited February 22, 20233 yr by mafiose13

10 hours ago, mafiose13 said:

But today with sunny weather I can confirm that we went from "premature float" to "never float".

I checked my patching, and it all checks out fine.

The only thing I can think of is that you have a long absorb time setting, possibly to compensate for the premature float bug. Please check the absorb time setting; it should be set to Aut. It's possible to set it to as much as 900 minutes (15 hours!). If you've never touched that setting (usually setting 32, misnamed "bulk charging time"), it should default to Aut, but it might have become overwritten for some reason.

10 hours ago, mafiose13 said:

This has got to be Not normal.

Definitely, and it is not good for your battery to be at the absorb voltage for too long. And you seem to have a moderately high absorb voltage, of 29.2 V (14.6 V per 12 V module). The default is 28.2 V (14.1 V per 12 V module). While the premature float bug is fixed, consider lowering the bulk/absorb voltage setting. Though hopefully it's still cold where you are, so that slightly higher battery voltages are normal and safe.

4 hours ago, Coulomb said:

Please check the absorb time setting; it should be set to Aut. It's possible to set it to as much as 900 minutes (15 hours!)

It's the first thing I thought about but we don't have this option on this model, the setting menu skips it. 

4 hours ago, Coulomb said:

And you seem to have a moderately high absorb voltage, of 29.2 V (14.6 V per 12 V module). The default is 28.2 V (14.1 V per 12 V)

The settings are straight off the battery (Eastman 240Ah Lead acid). They list Bulk 14.6v and 11.1v cutoff, 13.6v float charge. I usually set a higher cutoft at 12v. 

4 hours ago, Coulomb said:

Though hopefully it's still cold where you are, so that slightly higher battery voltages are normal and safe

It is cold out here, not freezing...7-15c temperatures. I think it's fine, I have temperature sensors on my set, never beyond 2-4 degrees from ambient temperature. 

4 hours ago, Coulomb said:

While the premature float bug is fixed...

I am under USE battery setting now. Im thinking about trying it under Fld to see if there is any difference, but the thing is i don't like that mode's default CV of 10.5v. I'll keep USE today, and try Fld tomorrow. 

Since the patch code checks out...any code thought/other ideas on your end? 

 

 

Edited February 23, 20233 yr by mafiose13

4 hours ago, mafiose13 said:

Since the patch code checks out...any code thought/other ideas on your end? 

I thought of sending a command to definitely disable timed absorb, since the code actually checks the variable and obeys it. But there doesn't seem to be such a command.

If you can, try sending a PF command, to change all the "user" settings to defaults (that should set the absorb time to AUT, even though there is no command or setting to change it). But it's possible that the display will not send it through. In that case you'd have to make the special cable and bypass the display. It starts getting hard by then. Maybe your monitoring software can reset to defaults; that makes it seem likely that the PF command should get through.

3 hours ago, Coulomb said:

I thought of sending a command to definitely disable timed absorb, since the code actually checks the variable and obeys it. But there doesn't seem to be such a command.

If you can, try sending a PF command, to change all the "user" settings to defaults (that should set the absorb time to AUT, even though there is no command or setting to change it). But it's possible that the display will not send it through. In that case you'd have to make the special cable and bypass the display. It starts getting hard by then. Maybe your monitoring software can reset to defaults; that makes it seem likely that the PF command should get through.

Tried switching back and forth User to Fld n vice versa, changed value for bulk, float, cv via  Watchpower sofware and through my phone App, reboot inverter between changes and after...None worked. It's stuck in the "never floating" mode. 

I left the inverter unattended at noon, came back at 6pm to no power at home, it was flashing 61 (communication lost-only available when battery setting as PYL per manual description), and mine was in Fld setting....so go figure!!! 🙄😬 (see attached)

So I powered down all breakers, inverter...restarted everything. Error 61 message disappeared...10 min later I get a buzzer n error F52 (low bus voltage). Another power down...didn't clear it. 

At this point I decided to reflash the un-patched 55.06. The inverter rebooted normally, no warnings yet it's been 2 hrs....and was immediately at float voltage 27.2v. 

All this bugginess & acting up...It's got to be something off with patch, it's beyond me. 

Edited February 23, 20233 yr by mafiose13

After some investigation, it looks like all this started to happen around 4:15-4:30pm when the sun is low (30 min pre sunset). It shouldn't have and it's never has. I can't put my finger on a cause.

I was running super low loads 200W or so, batteries full since noon.

Edited February 23, 20233 yr by mafiose13

6 hours ago, mafiose13 said:

It's got to be something off with patch,

I've thought of something else to check with the patch, now I seem to have lost some comments in the disassembly. I have an appointment all day today, so it will have to wait, sorry.

This is the hassle of patching code for machines you don't own. Sigh. My apologies in advance if it turns out to be my fault.

I've added a comment to the 255.06 patch download post to say it's deprecated for the time being.

I'm now thinking I must have some bad RAM pointers. If so, I'm amazed that it ran at all.

5 hours ago, Coulomb said:

I've thought of something else to check with the patch, now I seem to have lost some comments in the disassembly. I have an appointment all day today, so it will have to wait, sorry.

This is the hassle of patching code for machines you don't own. Sigh. My apologies in advance if it turns out to be my fault.

I've added a comment to the 255.06 patch download post to say it's deprecated for the time being.

I'm now thinking I must have some bad RAM pointers. If so, I'm amazed that it ran at all.

I volunteered this 1 unit for testing purposes and if that's what it takes...it is what it is. No worries 🙂👍...luckily as is it seems there was no permanent harm done, hopefully the next 24hr will tell. I'll keep a close eye. cheers 

15 hours ago, Coulomb said:

I've thought of something else to check with the patch,

Ok, that checks out as being correct.

Unfortunately, I found a really embarrassing bug where the resetting code won't take effect for the first 18 hours after powering up 😳.

So I'll do another version soon and fix that bug. However, it doesn't explain your weird happenings.

9 hours ago, mafiose13 said:

I volunteered this 1 unit for testing purposes and if that's what it takes...it is what it is.

That's a great spirit, thanks!

Edited February 24, 20233 yr by Coulomb

On 2023/02/24 at 8:57 PM, Coulomb said:

So I'll do another version soon and fix that bug.

@mafiose13, as promised.

This is patched firmware version 355.06, based on factory firmware 55.06 for the Axpert VM III non-twin, new models 1.5 kW or 4 kW (24V) or 6 kW (48V). The patches are modelled on patched firmware version 656.05 for the Axpert Max II 10kW:

* Threshold current is reduced from 0.50 A to 0.10 A (656.05 used 0.05 A, but there was a comment saying that it may be too low).

* The premature float bug is fixed.

* Every minute, if the PV voltage is below 175 V, the target PV voltage is reset to 200 V and the direction is forced to 1, meaning that we expect the PV power to increase by increasing panel voltage.

* The QVFW command returns 355.06, meaning that the display still shows the U1 version to be 55 06 (representing 55.06), but monitoring software should report the version as 00355.06, possibly without the leading zeroes.

* This version fixes a bug in patched firmware version 255.06, which prevented the resetting from happening for the first 18 hours after powering up.

I believe that all VM models use the same DSP processor ('28062), so there should be no problem with chipsets. Do not use this patched firmware with any other model, including other Axpert VM IIIs with rated power other than 1.5 kW, 4.0 kW, or 6.0 kW. However, this firmware can be used in VM IV models, if your existing firmware is or is based on 55.xx.

mafiose13 has reported that factory firmware version 55.06 fixed an issue he had with the AC output voltage not being set properly (ask for 230 V, get 215 V, for example). As far as I know, factory firmware version 55.06 is the latest firmware for non-twin Axpert VM IIIs (new type). [ Edit: It's still the only firmware I have for VM III and VM IV new (4kW and 6kW) as of January 2025. ]

Firmware upload instructions.

Use at your own risk.

dsp 355.06 patched VM III non-twin.zip

Edited January 24, 20251 yr by Coulomb

57 minutes ago, Coulomb said:

This version fixes a bug in patched firmware version 255.06, which prevented the resetting from happening for the first 18 hours after powering up

24hrs after reflashing the unpatched 55.06. All is functioning normally. No errors, no warnings, Battery reaches bulk n returns to float at 27v. Obviously what was happening was post patching. Any chance this bug above was causing this haywire behavior or something else missed? 

Ill flash this later tonight, test it tomorrow and report back, unless you figure out some other root for this.

On 2023/02/24 at 4:15 PM, Coulomb said:

This version fixes a bug in patched firmware version 255.06, which prevented the resetting from happening for the first 18 hours after powering up

Hey @Coulomb I flashed 355.06 last night, no errors or warnings since then. I also did not notice any foul play with MPPT stuck at any odd voltage so far, I'll keep observing as some of these issues started 1.5-2 days into the previous patch flash. However, the "Never floating" issue & staying stuck all day till sunset at bulk voltage 29.2v (even when charging amps drip to zero) has returned with this patched version as well, it had disappeared on the upatched version when i reflashed it in the interim between patch buggy 255.06 and this 355.06 . Somehow it's related to the patching code, placement...no idea

23 hours ago, mafiose13 said:

However, the "Never floating" issue & staying stuck all day till sunset at bulk voltage 29.2v (even when charging amps drip to zero) has returned with this patched version as well,

I wonder if fixing the premature float bug has allowed a latent bug to become visible. It could be very hard to find.

I can always reverse the premature float bug, but that seems backwards. I'll have a think about what this might be.

21 hours ago, Coulomb said:

I wonder if fixing the premature float bug has allowed a latent bug to become visible.

It's probably the case since it's patch-related somehow. 

Last night I had grid power overnight from 3-8am, so sunshine came in with battery already at 29.2V @ 4amps charge current, within 1 hr current dropped to 0amps, it stayed like this till sunset with occasional trickle charge to maintain bulk set voltage of 29.2v (over 8hrs). During the day i switched back n forth User to Fld and changed float from 26.5 to 27.2 a few times. Nothing triggered floating. The behavior seems to consider bulk voltage as float voltage. Out of fear of ruining the batteries I reflashed stock 55.06 again last night. Today, bulk/float works fine.  

21 hours ago, Coulomb said:

I can always reverse the premature float bug, but that seems backwards. I'll have a think about what this might be.

While this is figured out, Can we meanwhile do a version without the "premature float" patch code, this way I can test and tell you if that code is the source of the problem or not. At least we know which path to take in finding the solution. Unless you have other ideas. 

 

PS: Edit: I don't know if this clue makes any difference. when flashing the patched firmware the unit behaves normally on that day. It's only when it starts a new solar charging cycle (ie next morning) that it goes into the never floating mode. 

Edited February 27, 20233 yr by mafiose13

On 2023/02/27 at 5:59 PM, mafiose13 said:

Unless you have other ideas. 

Actually, I did. Someone at Voltronic must be at least dimly aware of the premature float bug, since they increased a counter threshold from 1500 (representing 30 seconds) to 30,000 (representing 10 minutes). That's the amount of time that the battery voltage must be over its threshold, and the battery current under its threshold, before the firmware will go to float. That's 30,000 times in a row, no exceptions! This is presumably to guarantee at least 10 minutes in the absorb stage, so as to not immediately ruin all lead acid batteries connected to Axpert inverters. A lazy "fix" for the premature float bug, it seems. In fully patched firmware, we put this counter threshold back to 1500, since I was finding it hard to get to float stage after we fixed the premature float bug many years ago. Of course, I completely forgot about that. Sigh.

So when I came across it again, I was sure that this was your problem too.

The only problem is, I was looking at an Axpert Max firmware; when I found the equivalent place in 55.06, I found that it this threshold was already at 1500! Arrgh, so close.

So now I'm back to not knowing why it won't go to float. That counter does get cleared to zero in some code that I haven't figured out yet, and it's one of those pieces of code that is different between VM and non-VM firmware. I looked through all the VM III firmware that I have, and none of them document this function. I even checked my one VM II firmware, and I don't have any VM IV firmware. So it will have to wait till I get time to analyse it.

In the meantime,

On 2023/02/27 at 5:59 PM, mafiose13 said:

Can we meanwhile do a version without the "premature float" patch code, this way I can test and tell you if that code is the source of the problem or not.

I'm pretty confident that this will "fix" it, but of course, the proof of the pudding is in the eating.

Attached is patched firmware version 455.06, based on factory firmware version 55.06 for the Axpert VM III 1.5/4/6 kW non-Twin models (not the earlier 5 kW etc models). The only difference from version 355.06 is that the premature float bug is temporarily unfixed (i.e. restored to factory original), as part of investigating the "never float" problem (see above discussions).

The usual caveats apply; use at your own risk.

[ Edit: A later post confirms that the premature float fix was not the problem. So the latest patched firmware for this model with the premature float bug fix is 355.06, available in this post. ]

 

dsp 455.06 patched VM III non-twin.zip

Edited May 22, 20233 yr by Coulomb